Skip to main content

Full text of "On the distribution of towns and villages in England"

See other formats


ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 511 



to the facilities granted by his Highness and with my folding canvas boats, I 
was pleased and happy in cirrying out at Ciieva del Drach one of my most satis- 
factory subterranean investigations. 

I found one of the largest underground lakes known in the world, which I 
named Lago Miramar, 570 feet long, 100 to 125 feet wide, 15 to 30 feet deep. 
The accompanying photographic illustration (taken with magnesium light in ten 
minutes) represents a corner of this lake, and shows better than any long descrip- 
tion what a marvel is this mysterious and so long unknown pond, black as night, 
but sparkling under magnesium wire with all the splendours of the diamond. The 
vaults and walls are covered with millions of sharp and thin stalactite needles ; 
the roof is supported at intervals hy stalagmitic columns resembling the Egyptian 
or Indian pillars of Karnac or Kailaja. All these concretions are pure white like 
ermine, without any spot of clay. It presents the greatest contrast to Gaping 
Ghyll, being a masterpiece of beauty, just as the Yorkshire abyss is of frightful 
magnitude, both stupefying to human eyes. 

And Dragon’s cave (li mile long) is not only a picturesque curiosity ; all its 
large and small lakes are on the same level as the sea, with which they communi- 
cate through narrow clefts ; they are half salt and fresh water, and a hydrological 
marvel, of which I will give elsewhere a fuller scientific account.* For geographers 
I only say here that this cavern is a mere sea-cave formed by the Mediterranean 
waves, and not hy an underground river, but a sea-cave of unusual size, unparalleled 
elsewhere, at least on European shores. It is said that California, Cuba, and Jamaica 
possess also grand caves of the same kind. When and by whom will they he 
scientifically explored ? 



ON THE DISTRIBUTION OF TOWNS AND VILLAGES 
IN ENGLAND. 

By GEO. G. CHISHOLM, M.A., B.Sc. 

II. Historical Aspects of the Question.! 

In examining this subject from the historical point of view, we begin, of course, with 
the Koman towns. In pre-Eoman times there were no towns. In the state of 
civilization reached by the Britons, there could be no use for them. So far as Caesar 
had the opportunity of observing, there was little cultivation in Britain in his time 



* Compfes Bendus de VAcadg’mie deg Sciences 14 Juin 1897, et Annuire du club alpine 
fraiifais, 1896 (in press). 

t Owing to a misunderstanding about the proof in the first of these papers 
(January, 1897), several misprints were left uncorrected, and should be corrected us 
follows : — 

Page 7t), line 3. For Barstone read Carstcne. 

„ 80, „ 3. For deepened read cheapened. 

„ 80, „ 17. For Buckmere read Cuckmere. 

,, 80, ., 25. For Steywing read Steyning. 

„ 81, note. For Gonglis read Gough’s. 

„ 82, line 2. For Whilton read Whitton. 

„ 82, ,, 9. For South read North. 

,, 86, ,, 1, For hillock read bullock. 



512 ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 



except in Kent, the inhabitants living on milk and flesh, and wearing skins for 
clothing. The name of “ town ” they gave, he says, to an inaccessible wood fortifled 
by a ditch and rampart, where they resorted to avoid invaders. 

The geographical interest attaching to the Eoman towns of Britain is all the 
greater because these towns appear to have been mostly abandoned after the 
Teutonic invasions of the fifth century. The organization of the Angles and Saxons 
hardly allowed a place for the existence of towns. Towns are mostly brought into 
existence as centres of trade, seats of manufacture, or places of administration ; but 
the remarkably self-contained village life of the Angles and Saxons enabled them 
for centuries to dispense both with manufacturing and trading towns. Even the 
administration of the villages was independent, and the only administrative centre 
required was a capital whence, as occasion demanded, the king might summon his 
immediate followers to attend him with their retainers. Hence the Eoman towns, 
where not still left in the hands of the Britons, were forsaken. The shell, or 
skeleton, of the town continued to exist in many cases till life was revived, if it 
was ever revived at all. Town life in England, according to Cunningham, did not 
begin again till the advent of the Danes, and was due to their trading instincts.* 
An interesting mark of this breach of continuity in town life, in the case of those 
Eoman towns which were revived and still exist, is found in the want of continuity 
of the town names. Very few old Eoman towns, or towns on the sites of old Eoman 
forts, in England are known by names directly descended, that is, derived by mere 
linguistic corruption, from those which they had in the Latin tongue.f 

Now, much of the geographical interest attaching to the history of Eoman 
towns arises from the hints we get therefrom as to the geographical conditions that 
may have favoured their existence in the situations where they are found or 
formerly existed. Where Eoman towns have continued to exist, or have been 
restored on the site originally occupied or on a site closely adjoining, there is at 
least a presumption that their existence on these sites is due to geographical con- 
ditions of an enduring kind. Where Eoman towns have passed away and left no 
modern representative, the inference is that, if their existence could ever be fairly 
ascribed to geographical conditions, these conditions have changed. 

Among Eoman towns or stations that have continued to exist or have been 
revived, it is natural to find a considerable number of seaports, including river- 
ports accessible to sea-going ships. Among these may be mentioned Londinium 
(London), Durobriv^ (Eochester), Dubrse (Dover), Clausentum (Southampton), 
Isca (Exeter), Glevum (Gloucester), Segontium (Carnarvon), Deva (Chester), Pons 



* ‘Growth of English Industry and Commerce: Early and Middle Ages,’ pp. 
88, 89. 

t How different from what we find in the names of ancient Gaul compared with 
those of modern France ! There most of the ancient Eoman names have come down in 
some form to the present day, and the towns themselves that were important in Roman 
times are more or less important still, and stand on the same sites. It is interesting, 
too, to note that the names that have thus persisted are often tribal names, and not the 
ofl[icial names that may have been given to the towns by the administration, but which 
were in many cases displaced by tribal names even in Eoman times. The names are 
often greatly corrupted ; still such changes as Lemovices, Limoges, Divio (nem), Dijon, 
are of the same nature as those which have changed ahbreviare into abr^ger, and are 
quite different from those which have taken place in names like Manchester and 
Lincoln, where new names have been formed by the combination of one element of 
the old name with a designation which the Anglo-Saxon settlers had learnt from the 
previous inhabitants. 



ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 513 



(Newcastle), Eburacum (York), Venta Icenorum (Norwich), and Camulodunum 
(Colchester). To these we may add Aqua3 Sulis (Bath), which, though it has long 
ceased to be a seaport, in all probability served all the purposes of a port for this 
district in Eoman times when no Bristol existed, and Luguvallium (Carlisle), which 
continued to be a seaport down to a comparatively recent date. 

Of the towns mentioned, Clausentum and Venta Icenorum do not appear to have 
occupied the exact site of their modern representatives, and the difference in 
situation, slight as it is in both cases, is not without interest. The exact site of 
the Eoman Clausentum is now occupied by the village of Bitterne, at the head, 
but on the east side, of the small estuary of the Itchen,* where Camden was 
shown ruins, walls, and ditches of an ancient castle about half a mile in circuit, 
and surrounded on three sides by the sea at flood tide, where, he adds, coins of 
Eoman emperors were frequently dug up. Through the care of the late Sir Stuart 
Macnaghten, the old Eoman boundaries of the station, which was an insular one, 
have been preserved, and can still be seen. The coins found here are chiefly of the 
later empire, those of Tetricus, one of “ the thirty tyrants,” being most numerous. 
A port at this spot would be the highest point that sea-going ships could reach 
on their way to Winchester, then the principal goal in this region. As traflic 
increased, the more extensive accommodation for loading and unloading on the 
peninsula between the Itchen and Test would give greater advantages to that site. 
Venta Icenorum appears to have occupied the site of the modern village of Caistor, 
which lies a little to the south of Norwich, on a small river called the Tese, probably 
navigable to this point in Eoman times, and would be the first navigable point 
reached in following the road to the Yare from the south. 

Besides Clausentum and Venta Icenorum, Portus Adurni (? Porchester), Tamara 
(near Tamerton, on the Plym), and Gariannonum (Burgh Castle, on the Yare), may 
also be mentioned as having their modern representatives as seaports at some little 
distance from the ancient towns. Whether Porchester is to be identified with 
Portus Adurni or not,t its name seems clearly to indicate the presence of some 
Eoman port here. It is now represented by Portsmouth, the foundation of which 
is ascribed by Camden to the removal of the inhabitants from Porchester to the 
island of Portsea, in consequence of “the gradual withdrawing of the sea.” At no 
period, however, can the harbour on which these ports stand have had anything 
like the commercial importance of Southampton, which is much better situated for 
communication with the interior, and does not suft’er from the disadvantage of the 
strongly flowing and ebbing tides, which at times render the navigation of the 
entrance to the other port extremely difficult. The commodiousness of the harbour as 



* The very plausible derivation given by Camden for the name Clausentum would 
seem to militate against this identification. Camden suggests that Clausentum means 
the harbour of the Entum, or Anton, claudh meaning in “ British ” a harbour formed by 
artificial banks of earth (Pugh’s AVelsh dictionary defines Claicdd as “ a dike, ditch, 
or trench, a pit or quarry ; an embankment or wall thrown up ; ” the primary meaning 
evidently being “something made by digging ”). The dh, or Welsh dd, pronounced 
like the th in this, would readily enough become an s ( 2 ) in the mouths of those who 
could not pronounce that sound. But Anton is the name of the river on the west side 
of the small peninsula on which Southampton stands. The name may very well have 
been given also, however, to what we call Southampton Water, and hence the name 
“port of Anton” would be suitable for any port at the head of that inlet. 

t “ Porchester Fort, possibly Portus Adurni.” — F. Ilaverfield, in the list of places 
contained in the map of Eoman Britain in the ‘ Historical Atlas of Modern Europe ’ 
(Oxford ; Clarendon Press, 1896). See also Archaeological Journal, vol. xlvi. p. 68. 

No. V. — November, 1897.1 2 m 



514 ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 



a naval station is the main cause of the importance of Portsmouth, Modern railway 
facilities were required to give it the commercial value which it now also possesses. 

Tamara, now represented by Plymouth, owes its displacement as a seaport to the 
increased size of vessels. When smaller sea-going vessels were in use, it was, of 
course, important that they should penetrate as far inland as they could. The 
displacement of Gariannonum, now represented hy Yarmouth, is, on the other hand, 
probably due to a change in the coast-line. In Eoman times, the present site of 
Burgh Castle appears to have been at the mouth of the Yare. Yarmouth, which 
dates from Saxon times, stands on a spit of blown sand, which holds back recent 
alluvium of the Yare and Bure. 

Among Eoman seaports which have not survived, at least as such, may be 
mentioned Portus Lemanis (Lympne), already referred to (vol, ix. p. 85) ; Eegulbium 
(Reculver) and Rutupise (Richborough), both of which have lost their importance 
through the alluvial deposits which have filled up the Wantsum and the mouth of 
the Stour ; Anderida (Pevensey), which has also been cut off from the sea by 
alluvial deposits ; Regnum (Chichester) ; Branodunum (Brancaster) ; and Vainona 
(Wainfleet). Whether Regnum is, strictly speaking, to be included among the 
Roman seaports of Britain is perhaps doubtful. No doubt part of its importance 
was due to the fact that boats of light draught can navigate Chichester harbour (the 
inlet to the east of Hay ling island) ; but it does not follow that they could reach the 
site of the city itself, the geographical importance of which is partly determined by 
other circumstances (see vol. ix. pp. 76, 77). Camden notes that the city “ would have 
arisen to great splendour had not the harbour been at too great distance and incon- 
venient,” and he adds that the inhabitants were then about to dig a new canal. 
The Lavant, the small stream on which the city stands, is useless for navigation ; 
and the canal, which now admits barges of 4 feet draught to the city, follows a 
different course. 

Branodonum is only known as one of the forts of the Saxon shore, but it is 
probably to be regarded as an ancient seaport, for there is no other place that it 
could have served as a defence for, so that it must have been erected to defend a 
landing-place from the sea. It is now quite cut off from the sea by the marshy 
alluvium already mentioned (vol. ix. pp. 76, 77), as here lining the base of the chalk 
cliffs of Norfolk. 

Wainfleet, on the Lincolnshire shore of the Wash, is identified by Dr. Stukeley 
with Vainona (or rather with the Navione of Ravenna, of which Vainona is a 
corrected reading) ; and Mr. Skertchley, making the same identification, believes 
that here in Roman times was the main mouth of the Witham.* Leland (1549) 
speaks of Wainfleet at that time as a pretty market town on a creek near to the 
sea, with small vessels belonging to it ; and this creek, he says, after entering a 
considerable way inland, sent out branches on either side into the fens. 

Of Roman towns that stood on navigable rivers, the following may be mentioned 
as still retaining some of their ancient importance : Durovernum (Canterbury), 
CcEsaromagus (Chelmsford), Camboritum (Cambridge, or rather Grantchester, the 
predecessor of Cambridge), Durolipons (Godmanchester, opposite Huntingdon), 
Ratae (Leicester), Lindum (Lincoln), Danum (Doncaster), Legeolium (Castleford), 
Calcaria (Tadcaster). Most of these were probably either at the head of naviga- 
tion on the rivers on which they stand, or at fords, or both. The Stour was 
navigable to Canterbury down to comparatively recent times. The name of 
Cambridge, and that of its ancient representative, show that the bridge across the 



* “Geology of theFenland.” By S. B. J. Skertchley (‘ Memoirs of theGeol. Survey 
of Great Britain and Ireland, 1877 ’), p. 15. 



ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 515 



river here was preceded by a ford, ritum — being the Latinized form of the Welsh 
rJiyd, “ a ford.” The name Castleford points to a similar origin for the town and its 
ancient predecessor. Leicester was long at the head of navigation of the Soar, the 
navigation being now continued beyond the town by canal. Tadcaster is still the 
head of navigation on the Wharfe. The importance of Lincoln as, on account of 
its position at a well-marked break in the Lincolnshire limestone, a point of con- 
vergence of roads, has often been remarked on. In Roman times, and for many 
hundreds of years afterwards, it was nerhaps quite as important as a meeting-place 
of waterways. The Witham below Lincoln is still a waterway to the Wash for 
boats of 5 feet draught, and this line is continued west of Lincoln to the Trent at 
Torksey by the Foss dyke. The latter navigation at least was much more im- 
portant formerly than now. According to Dr. Stukeley, Torksey occupies the 
site of a walled Roman town, and the castle existing in his day was founded, he 
says, on the old Roman granarv, “ which was much like Colchester castle, with 
circular towers at the corners.” Southwards, Lincoln communicated with the 
Welland by the Car-dyke (that is, fen-dyke), the Roman navigable trench by 
which the waters from the higher grounds on the west were intercepted and thus 
preyented from drowning the low fenlands to the east. It began on the Welland 
to the west of Ermine Street, and, passing to the east side of that highway at 
Cateshridge. continued northwards by a nearly parallel course till it ultimately 
joined the Witham. The ditch still survives, though no longer navigable. In 
Roman times, and probably for long after, it must have added considerably to the 
importance both of Lincoln and Torksey. In Domesday Torksey is mentioned as 
a place with “ two hundred burgesses enjoying not a few privileges, upon condition 
they should conduct the king’s ambassadors when they come this way in their 
boats down the Trent, and bring them as far as York.” * In Norman times, Lincoln 
is described by William of Malmesbury as one of the most populous cities in 
England, and a mart for commodities brought both by land and water. By 
Edward III. it was made a staple for wool, leather, lead, etc. ; but two centuries 
later Camden says it is incredible how it had declined by age, only eighteen churches 
then remaining of the fifty which it had held in Edward III.’s time. By that time 
the Poss-dyke had got choked up. Bishop Atwater (1514-21) had begun to cleanse 
it in the hope of bringing yessels to Lincoln, but died before achieying his purpose.f 
Of Roman river-towns which, like Torksey, have decayed without recovery, 
four — Isurium (Aldborough), Durobrivse (Castor), Margidunum (East Bridgford), 
and Uriconium t (Wroxeter) — are worthy of special notice. Next to York, Isurium, 
which stood on the Ure a short distance above its confluence with the Swale, was 
the most important city in the valley of the Ouse. It is close to the terminus 
of river-navigation at the present day, the navigation being now continued to 
Ripon by canal ; but this in itself could not have accounted for the growth of 
a town at this point, as the Swale, or eastern headwater of the Ouse, continues 
to be navigable a good deal further north. Probably the adoption of the site 
was due to the fact that, in proceeding northwards on the right side of the 
river by the direct road from Tadcaster, this was the first point at which the 
river could be conveniently forded, the road then proceeding beyond this point, 
between the Ure and the Swale, to Cataractonium (Catterick Bridge), where the 
latter river was crossed. After its decay, Isurium never revived, and such im- 
portance as it may have possessed as a centre of trade seems to have passed over to 



* Camden’s ‘Britannia,’ Gough’s edit., vol. ii. p. 227. 
t Leland, vol. i. p. 32. 

X OrYriconium; also found as Viroconium. 



2 M 2 



A. 



A. 



516 ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 

Ripon, the town which grew up round the ahhey founded by Archbishop Wilfrid 
of York, in the infancy of the English Church. 

As the importance of Isurium has passed over to Ripon, so that of DurohrivEe 
(“the passage of the river” Nen) has passed over to the city of Peterborough, 
which owes its origin to a monastery built in the early part of the seventh century 
by Peada and Wolpher, sons of Penda, the first Christian king of Mercia, a little 
to the east of the old Roman town, the site of which is still commemorated in the 
names of two villages. Castor, on the north or Northamptonshire (once the Danish) 
side of the river, and Chesterton, on the south, Huntingdonshire, or Anglian side 
of the river. Similarly, the importance of Margidunum, whatever that may have 
been, has passed over to Nottingham. The site where Roman remains have been 
found identified with those of Margidunum is on the Trent a few miles north-east 
of Nottingham, and the place may have had a certain importance in connection 
with the Trent navigation, since it is the first point at which the Fosse Way, 
coming from the south-west, approaches the river ; but there are no evidences of 
any important crossing-place here in Roman times as there was in later days near 
Nottingham. Probably the forests of western Notts (Sherwood, etc.) extended 
further east in Roman times, and hence the most convenient communication with 
the north was first north-east to Lincoln, and then north-west between the forests 
of the present Nottingham and the marshes of the lower Trent and Don by 
Torksey, Doncaster, and Castleford to Tadcaster and York. 

The remaining Roman towns that it may be interesting to consider are best 
noticed in connection with the great Roman roads. The great north-west road is 
that which came to be known as Watling Street. If we take it as beginning at 
Dover or Richborough, it passed by Canterbury and Rochester to London, then by 
Sulloniacse (Brockley Hill), Verulamium (St. Albans), thence to the south point 
of the modern county of Leicester, from which point it forms the whole of the 
south-west boundary of that county (the only county boundary composed of 
nearly mathematically straight lines in England). On leaving Leicestershire, it 
first continues in a north-westerly, then in a westerly, direction to the north of the 
ancient forests between the Avon and the Severn, passing Letocetum (Wall), 
Pennocrucium (Penkridge), and touching the Severn at Uriconium (Wroxeter). 
From that point it probably ran north to Chester. 

Important as this road undoubtedly was for centuries, it is somewhat remark- 
able that there is no town on it of any importance north of the Thames except 
St. Albans and the terminal city on the Dee. St. Albans, though always notable 
in English history, is of small consequence compared with what it was in Roman 
times, when it was the rival of London in importance and dignity, though not, of 
course, in commerce. If Letocetum had any importance in Roman times, that 
importance has passed to Lichfield, which is situated 2 or 3 miles to the 
north-east. Uriconium was certainly an important Roman town. Its precise 
position seems to have been determined by the fact that the Severn is here fordable, 
which it is not lower down. After the destruction of this towm by the Danes, it 
may be said to have been replaced by Shrewsbury, the town which arose a few 
miles higher up, on a red hill t nearly encircled by the Severn — a position 
accordingly peculiarly well adapted for defence. 

This Roman road was crossed by another called the Fosse Way, running from 
south-west to north-east. It no doubt branched ofi' from some point on the great 



* Etocetum, hitherto the usual, is, it seems, an incorrect spelling. See the list 
already referred to in the ‘ Historical Atlas of Europe.’ 
t Lower Permian (Rothliegendes). 



ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 517 



south-western road to Exeter, but only begins to be traced in the neighbourhood 
of Ilchester, whence it passed north-east to Aquee Sulis (Bath), Corinium (Ciren- 
cester), and that point in the south-west boundary of Leicestershire where a slight 
change is made in the direction of that boundary (from north-west by north to 
north-west by west). At this point it crossed Watling Street, and beyond that 
road its north-easterly direction was continued by Leicester to Lincoln. 

On the Fosse Way, the only Eoman towns besides Lincoln which have retained 
any importance are Bath, Cirencester, and Leicester. Bath still derives importance 
from the hot springs, to which it owes its present, as it owed its Roman, name, and 
by which in ancient times it gained celebrity. Cirencester, though still a town 
with several thousand inhabitants, no longer has the relative importance which it 
must have had in Roman times, when it was the point of convergence of several 
important roads, the Fosse Way being here crossed by the road from Silchester to 
Gloucester. It was probably the fact of the Soar navigation ending here that deter- 
mined the revival of Leicester on its ancient Roman site. 

An important Roman town might have been expected to exist at the point on 
the Leicestershire boundary already mentioned as the crossing-place of Watling 
Street and the Fosse Way. At present there are only one or two small villages in 
the neighbourhood of this spot, which still bears the name of High Cross, “ from a 
cross,” says Camden, “that anciently stood there on a high ground, now succeeded 
by a higher post with its props;” but the people round about stated that “here was 
a most flourishing city named Cleycester, which had a senate of its own, and of 
which Cleybrook (Claybrooke), about a mile off, was part, and that on both sides of the 
road foundations of hewn stone lie under the furrows, and Roman coins are very 
often ploughed up.” This city he identified with the Bennon®, or Venonse, of the 
Itinerary of Antonine, and the identification has been generally followed. 

Dr. Guest contends that the southern portion of the Ermine Street, or great 
north road leading out of London, was not of Eoman construction, basing his 
belief on the absence of Roman stations, villas, and burying-grounds on the part of 
the road between London and Huntingdon (or Godmanchester — Durolipons), and 
on the fact that of the three Antonine itinera leading to London, two first run 
down Watling Street to the Fosse Way, and thence from Venonje north-west, and 
one runs first to Colchester, and then by Cambridge to Huntingdon. If Dr. Guest 
is correct, we may conclude that the alluvium of the Lee, and the forests on both 
sides of that river, presented obstacles in the way of road-making which it was 
not worth while to overcome under the conditions then obtaining. The route 
north-eastwards led to the most populous and richest part of the country by a 
succession of places where deposits of gravel and sand (Ilford, Romford, Ingate- 
stone) and pebble-beds (Brentwood) afforded sites for towns and villages, and to 
some extent, no doubt, road-making material. The stations between Colchester 
and Cambridge on the iter leading that way to Lincoln are not Identified. The next 
beyond Godmanchester is Durobrivai, already mentioned as identified with Castor, 
near Peterborough; and the only other till Lincoln is reached is Causennae, 
identified with great probability with Ancaster, now only a small village a few 
miles north-east of Grantham. On this route, also, accordingly the only places 
identified still of importance are those already mentioned as sea or river ports. 

It is the same north of York. As far as Newcastle, not a single Roman station 
in the north has been identified with a place which is now of any importance. 
West of the Pennine chain, Carlisle, which has always had some importance as 
the centre of the New Red Sandstone (Triassic red Permian) basin of the Eden> 
seems to be the only place of importance now which was also important in Roman 
times until we come to the Dee. It is noteworthy that, at least in the time of the 



518 ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 



Itineraries of Antonine, the main line of communication with the far north, whether 
towards the east or west end of the wall, was by the east of England. The fifth 
iter, which gives one route from London to Carlisle, passes from the basin of the 
Tees to that of the Eden by the route now followed by the North-Eastern Railway. 
Bowes, the village which gives name to the station at the head of the valley on 
the east side of the upland part of the railway, occupies the site of the Roman 
station of Lavatrae on this route. The Roman station of Verteras, near the head of 
the valley on the west side of the Pennine chain, occupied the site of the modern 
village of Brough, a little to the east of the northern branch of the North-Eastern 
line on this side. From Brough the Roman road passed down the valley to Carlisle. 
There was one intermediate station, and it may be remarked that this station stood 
not exactly on the site of the modern Penrith, the point on which one descends 
directly in coming from Shap Fell, by the route by the London and North-Western 
Railway and the high-road from Lancashire, but at the place where the road from 
the north-east crosses the Eamont just before reaching Penrith, namely, at Brougham 
Castle, whose name probably preserves that of the Roman station Brocavum. 

There are two itinera giving routes from the north to Manchester and thence 
southwards, and both of these branch off from the route just spokeu of. One of 
these (iter II.) takes us from Carlisle to York, and then proceeds south-west by 
Tadcaster, and crosses the Pennine chain a second time before reaching Manchester. 
One intermediate station, Cambodunum, is mentioned, and this is now generally 
identified with Slack, which is now not even a hamlet, but a mere site about 
miles west by north in a direct line from the London and North-Western station 
at Huddersfield. The other route (Iter X.) is a very obscure one, and was made 
all the more obscure by the inconsistencies and various readings of the texts that 
Camden, Horsley, Reynolds, and other English antiquaries, had to deal with. A 
very plausible interpretation of this route, based on the revised text of Parthey and 
Pinder, is given by Mr. W. Thompson Watkin in vol. xxviii. of the Archaeological 
Journal. According to this interpretation. Iter X. crosses Shap Fell a little to the 
east of the north road from Kendal, and, after crossing the Eden valley, is continued 
by the Maiden Way referred to by Scott in ‘ Guy Mannering ’ (ch. xxiii.) in the 
direction of the wall ; but this view seems to be only partially adopted by Mr. 
Haverfield in his map of Roman Britain. Like Mr. Watkin, Mr. Haverfield appears 
to identify, though doubtfully, Coccium (an unimportant station) with Wigan on his 
map, but he does not enter it in his list ; and Bremetannacum and Galacum are 
identified by both with Ribchester and Overborough respectively, the former now a 
mere village, though probably a considerable seaport in Roman times ; the latter a 
site 2 or 3 miles south of Kirkby Lonsdale, about the place where the Lune enters 
Lancashire, now represented not even by a hamlet. Beyond this point Mr. Haver- 
field looks upon the route of the iter as dubious, though some Roman road is 
marked by him as following the remainder of the route marked out by Mr. Watkin. 
But on this part of the route also there was no station now of any importance, 
unless wo may regard the modern lead-mining town of Alston as representing an 
ancient Roman lead-mining centre under the guardianship of a station (a mile or 
two to the north) at Whitley Castle, which is identified by Mr. Watkin with 
Glanoventa, and has been identified by others with other stations. All interpreta- 
tions of this iter agree in identifying Mancunium with Manchester, but there is no 
evidence that Manchester itself was a place of any importance in Roman times. 
The Roman remains found there are scanty, and its site in Roman times had no 
advantages. Marshes, forests, and sandy wastes seem to have then rendered south- 
west Lancashire of little value. The site seems to have been determined by the 
“ hard rock of stone ” which Leland speaks of on the banks of the Irwell, a good 



ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 519 



foundation for a fort. The Irwell afforded power at an early date for mills, but 
it was not till after 1720 that the “ vadys and rokkys ” that obstructed its naviga- 
tion were removed, and the river made navigable to the city by means of weirs and 
locks. 

In the south-west of England, two inland towns besides Bath have maintamed 
more or less importance since Eoman times. These are Winchester and Dorchester, 
and along with these may be mentioned Salisbury, the modern representative of 
Sorbiodunum, though on a slightly different site. The natural features, all very 
similar, to which these three towns owe their importance have been already re- 
ferred to (see vol. ix. pp. 80, 81). All the other stations of the itineraries in 
south-western England are identified with places now quite insignificant, unless 
Horsley’s identification of Vindomum with Farnham be accepted; but one of them, 
Calleva Atrebatum, now Silchester, must in Eoman times have been a centre of 
very considerable importance. Its Eoman remains cover an extent of 80 acres, 
and in Eoman times it was a meeting-place of roads from London, Cirencester, 
Salisbury (Old Sarum), and Winchester. 

In Wales and Monmouth and the adjoining counties, three principal routes are 
given in the Itineraries — one along the southern maritime tract to Carmarthen, one 
along the northern maritime tract to the Conway river, and one running from south 
to north, east of the Welsh mountains, connecting the first route with Wroxeter 
and Chester. The Eoman ferry across the Severn appears to have been about the 
place where the estuary begins to expand, in the neighbourhood of the present 
tunnel. The landing-place on the English side is not definitely ascertained, but on 
the Welsh side it must have been about Portskewet, for the first station on that 
side, Venta-Silurum, a very important one, has been clearly identified with Caer 
Went, on a height immediately above that village. From that station the southern 
road running west appears to have touched four seaports, all of which are seaports at 
the present day, or have a representative immediately adjacent. The first of these 
seaports was Isca Silurum, remains of which are found at Caerleon, on theUsk, from 
which the station took its name, the seaport on which, however, is no longer at 
Caerleon, but at Newport, a short distance lower down. The other three seaports 
are Nidum (Neath), Leucarum (Loughor, or, in Welsh spelling, Llwchwr), and 
Maridunum (Carmarthen). In the north the road took a more inland course, its 
line being apparently determined, not by the ports, but by the fords. Only two 
stations west of Chester are mentioned in the Itineraries, Varm (Bodfari) and 
Conovium (the “ Conway ” station, placed by Horsley at Caerhun). Probably this 
road was afterwards continued to Carnarvon (Segontiacum), which is mentioned in 
the Notitia, but not in the Itineraries or by Ptolemy. 

The road leading from the south to Wroxeter started at Caerleon, and first 
followed the Usk valley. The intermediate stations on it are identified with Usk, 
Abergavenny, Kenchester (north-west of Hereford), and Leintwardine (crossing-place 
of the Teme in north-west of county of Hereford). Two of these have still a 
certain amount of at least relative importance arising from their situation. Usk, in 
Eoman times as now, was the starting-point of a road, now a railway, ascending the 
valley of the small stream that here joins the Usk river, and passing by Monmouth, 
Eoss, and just north of the Forest of Dean to Gloucester. Abergavenny stands at 
a still better-marked node, the point of convergence of four valleys, the upper and 
lower Usk forming two of these (running respectively north-west and south), that 
followed by the railway to Merthyr Tydvil forming a third, and that by the Great 
Western line to Hereford the fourth. The other two stations have been replaced by 
modern towns at a greater or less distance from the ancient, their sites being now 
occupied only by insignificant villages. 



520 ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 



So much with regard to the towns or more important stations which did exist 
in Eoman times. It is in some respects even more instructive to consider shortly 
some of those which did not then exist. And from this point of view there is, 
perhaps, nothing more striking than the scanty indications of Eoman settlement on 
the three great rivers, the Thames, Severn, and Trent, above the limits reached by 
sea-going vessels. True it is that Eoman stations, as we have already seen, are 
mentioned in the Itineraries at points far from the mouth on all these rivers; but all 
of them seem to have been mere crossing-places, and, except in the case of Wroxeter, 
there are no remains to show that they were of any great importance. Pontes, on 
the Thames, whether at Old Windsor or Staines, was merely the crossing-place on 
the road to Silchester. The name of the station. Ad Pontem, on the Fosse Way, 
together with the distances mentioned in the fifth Itinerary, makes it extremely 
probable that there must have been a station at a crossing- place of the Trent about 
Farndon, just above the islands enclosed by the two arms of the river, on one of 
which now stands Newark. The station must have been on the right bank, as 
there is no mention or indication of any other crossing of the Trent on the way to 
Lincoln, and no trace of any Eoman road running north on the left hank of the 
river. There must, of course, have been some population on that bank of the river, 
but that population seems to have been very scanty. The absence of all traces of the 
station at which the river was here crossed seems to justify the inference that there 
was no such population on the left bank as to give to the station any great impor- 
tance. At Wroxeter it was different, but the difference was not due to any addi- 
tional value given to the site by the navigation of the Severn, but to the fact that 
the station at this ford was the point of convergence of roads from the south-east 
(London), the north (Chester), and the south (Caerleon and the Severn estuary). 

Of course we have other evidence of the existence of Eoman towns and settle- 
ments than the fact of their being recorded by one or other ancient writer. There 
is the evidence of names and remains. There are, it is said, 119 -chesters or -cesters, 
or some other form of the Eoman castra, in Domesday, and it is probably fair to 
presume that all of these represent Eoman towns. There are none of these, how- 
ever, in the course of the Trent. The equivalents' of the Eoman strata are more 
doubtful signs of a Eoman origin, but there is no name of this root met with in 
ascending the river till we come to Stretton, near the mouth of the Dove (the 
valley containing Uttoxeter). On the Thames between Cirencester and London, the 
only name indicating a Eoman origin is Dorchester, and the Eoman remains found 
here confirm the indication. Possibly, seeing that it is not mentioned by any 
ancient writer, the place is of late Eoman origin. In any case, its existence on the 
banks of the Thames is no sign of extensive settlement on these banks, for it also 
was only a crossing-place — from the vale of Aylesbury, north of the Chiltern hills, 
to that of the Ock, in Berkshire. There are several “fords ” about this part of the 
river, and the name Shillingford, exactly at the mouth of the Thame, would 
indicate that there was once a ford close to Dorchester. Eoman remains have also 
been found at Sinodun hill, on the opposite bank. 

Between Wroxeter and Gloucester there is also only one -cester — Worcester. 
Here, however, there are no Eoman remains, though an old parchment, Camden 
tells us, boasted that it once had Eoman walls. Probably it too was of late Eoman 
origin, and if Prof. 'I'horold Eogers was right in saying (I don’t know on what 
authority, except perhaps on the presumption that Droitwich is the Salinse of 
the anonymous writer of Eavenna) that the Eomans undoubtedly made use 
of the Worcestershire salt deposits, it is possible that the salt was conveyed down 
the Severn from Worcester. 

All these indications of Eoman occupation on the banks of the three great rivers 



ON THE DISTRIBUTION OF TOWNS AND VILLAGES IN ENGLAND. 521 



of England are, however, very scanty, and the absence of others would seem to 
justify the conclusion that in Eoman times the valleys of these rivers were very 
thinly peopled. In all probability they were still, for the most part, unreclaimed 
marsh and forest. 

If that was so, that single fact implies that the conditions under which towns, 
having practically ceased to exist in England, arose again after four or five centuries 
of village life, were to a large extent different from those under which they had 
been created by the Komans. During this long interval, an interval as long as 
that between the Wars of the Kosesand the present time, but of which our only 
record is a story, and that imperfect, of wars between Angles, Jutes or Saxons, and 
British, between Saxons and Saxons or Angles, and between English and Danes, 
the great work of reclaiming the land, making it fit for agricultural settlement, 
must have gone on incessantly. Great numbers of the villages to which continental 
invaders after the departure of the Romans gave their names must have existed in 
Roman times, but multitudes of others must have received their names from these 
settlers through the original right of first occupation. And it is to be noted that 
when towns did come into being under these circumstances, they would grow ; they 
would not be created. They would serve in some way local convenience, and 
increase from villages to towns in proportion as they did so. On the other hand, 
the Roman towns or more important stations on the great roads appear to have 
been established merely with reference to military and administrative convenience, 
and without any special regard to the trade requirements of the districts in which 
they were situated. Some of the more important towns, such as London, Lincoln, 
York, Winchester, Bath, Gloucester, and others, had their sites determined by 
natural cofiditions which have made themselves good throughout the history of 
this island. These were almost of necessity fixed points in any network of roads 
constructed for military purposes. But between such points the roads were laid 
with as much directness as the state of the country permitted, and no deviations 
seem to have been made, even for the sake of passing through important mining 
districts. Though pigs of lead with Roman inscriptions, dated in the sixth year 
after the invasion of Claudius, have been found in or near the Mendip hills, showing 
that the lead-mines here were worked from the very beginning of the Roman occu- 
pation, the Fosse Way is laid so as to run to the east of those hills, the mines in 
which were served by a mere vicinal road. So also, if Ariconium in the thirteenth 
Iter is rightly identified with Ross, the distance to Gloucester given in the Itine- 
raries does not admit of any deviation into the mining district of the Forest of Dean, 
the produce of which must also have been carried along vicinal roads, or possibly on 
pack-animals by mere tracks. 

A few important places having thus determined the direction of the main roads, 
the intermediate stations — which became, in some cases at least, the nuclei of con- 
siderable towns — had their sites necessarily determined by the direction of the 
roads, and only in a secondary degree by local circumstances, and these determining 
local circumstances were regarded rather from the military and administrative than 
the commercial point of view. Hence it is that, while many of the more important 
Roman towns have their modern representatives on the same site or a site imme- 
diately adjoining, many others have utterly disappeared, or have given place only 
to insignificant villages. 

Nevertheless, there may have been reasons in certain cases why new towns 
should grow up under new conditions on or near the sites of some of these inter- 
mediate stations. This is notably the case with ford towns or stations. The 
position of intermediate Roman stations on rivers seems to have been determined 
rather with reference to the convenience of crossing the river by fords or bridges